Research

IntervalCoach learns how each athlete responds to training, night by night. Pooled across everyone, that data answers questions about training and recovery, and the answers change how the coach behaves.

Studies

  1. Individual recovery curves, now fitted from every morning

    What we found
    Modelling how the effects of consecutive sessions overlap turns about a hundred mornings per athlete into usable evidence, instead of the four or five a clean gap between sessions allows. On mornings the model had not seen, the new curve is more accurate for about six athletes in ten.
    What we changed
    Recovery depth is now fitted individually for every athlete, and the recovery rate for every athlete whose data supports one. Where it does not, the population value is used and labelled as such.
    Read the study →

  2. Can you see illness coming? What the overnight data shows

    What we found
    Recovery outlasts the illness. Overnight resting heart rate and HRV are most disrupted around four days after a sick spell ends and take about two weeks to settle, and how many days you were ill is a poor guide to when you are ready.
    What we changed
    The return to training after illness now lasts at least five days and holds the easing until your resting heart rate and HRV are back at your own normal, whatever the calendar says.
    Read the study →

  3. Do fitter athletes recover faster? What our recovery curves show

    What we found
    Fitness barely changes how fast you recover. It changes how hard a session hits: pooled across hard session types, the fittest athletes take roughly 40% less of an overnight hit than the least fit, while the recovery half-life stays near a day in every fitness band.
    What we changed
    Your fitness now scales how deep each session's hit is in your recovery curve, while the recovery rate stays tied to the type of session. That is why a beginner's plan leaves more room around hard days, and why that room closes as they get fitter.
    Read the study →

How we run them

Measured against your own normal

Every signal is read as a deviation from that athlete’s own baseline, so athletes with very different resting values can be pooled.

Every study changes the product

A study ends with what we changed because of it. That is the reason we run them.

Re-run as the data grows

Each study is repeated on fresh data before it is published, and updated when a later run moves the result.

Observational, and labelled that way

These read our own athletes’ data. They are not peer-reviewed trials, and they describe our population.